Potential additions to period search techniques

[fcotizelati]

Hello,

(This is not really an issue!)

I recently came across this paper: A Method for Estimating X-Ray Pulsar Period and Pulse Time Delay: Applying the Improved Zn,bin²-test to Complex Profiles. The technique described seems particularly interesting for scenarios with low photon counts and complex pulse profiles.

The authors introduce two techniques:

1. Instead of using a fixed number of harmonics for the period search, the method dynamically selects the optimal number of harmonics based on the photon count in the observation using the H-test.

2. Instead of using the bin midpoint as the representative phase when folding data, the proposed approach segments each bin by "equal photon intensity". They mention that this results in a more accurate phase estimation because in this way the folded profile should better reflect the true pulse shape, even in the presence of sharp transitions.

I am wondering if you think these methods are interesting and/or there are already any plans to integrate similar adaptive techniques in Stingray’s period search tools?

[matteobachetti]

@fcotizelati very interesting method, and I'd say well in scope for stingray. If you have code ready, feel free to contribute. In any case I'll take a look

[fcotizelati]

Hi @matteobachetti,

Thank you for the positive feedback. I have just opened a pull request where I implemented some new functions in the search.py module. I named the relevant module z_n_search_adaptive (not sure if it's a suitable name) and it takes the same arguments as z_n_search, except for the number of harmonic components assumed for the signal, which is set by the H-test.

I ran a test assuming a Crab-like profile shape, and I attach you a comparison of the periodograms obtained between the currently implemented method (assuming two harmonics) and the new method.

I hope that the procedure I followed to make contribution is correct, and I would appreciate your feedback on the code written so far. Thanks!